CN104915672A - Rectangular building extraction method and system based on high-resolution remote sensing image - Google Patents
Rectangular building extraction method and system based on high-resolution remote sensing image Download PDFInfo
- Publication number
- CN104915672A CN104915672A CN201410093123.4A CN201410093123A CN104915672A CN 104915672 A CN104915672 A CN 104915672A CN 201410093123 A CN201410093123 A CN 201410093123A CN 104915672 A CN104915672 A CN 104915672A
- Authority
- CN
- China
- Prior art keywords
- angle
- line segment
- shaped
- rectangular
- straight
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 40
- 239000013589 supplement Substances 0.000 claims abstract description 8
- 238000012805 post-processing Methods 0.000 claims abstract description 7
- 239000000284 extract Substances 0.000 claims description 26
- 238000001514 detection method Methods 0.000 claims description 14
- 239000000203 mixture Substances 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 4
- 230000015572 biosynthetic process Effects 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000003786 synthesis reaction Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 230000008878 coupling Effects 0.000 claims description 2
- 238000010168 coupling process Methods 0.000 claims description 2
- 238000005859 coupling reaction Methods 0.000 claims description 2
- 238000012946 outsourcing Methods 0.000 claims description 2
- 239000000463 material Substances 0.000 abstract description 3
- 239000003086 colorant Substances 0.000 abstract description 2
- 238000007781 pre-processing Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 241001212149 Cathetus Species 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 241000165780 Preussia/Sporomiella species complex Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000010197 meta-analysis Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- VMXUWOKSQNHOCA-UKTHLTGXSA-N ranitidine Chemical compound [O-][N+](=O)\C=C(/NC)NCCSCC1=CC=C(CN(C)C)O1 VMXUWOKSQNHOCA-UKTHLTGXSA-N 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000013341 scale-up Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Abstract
The invention relates to a rectangular building extraction method and a system based on a high-resolution remote sensing image. Straight line segment features in the remote sensing image are firstly extracted, right angle structures are extracted then by using a straight line segment extraction result, and the right angle structures are organized to form rectangular structure elements; as for isolated right angle structures and U-type structures which can not be combined into complete rectangular structures, rectangular elastic contours are adopted to supplement un-extracted boundaries by using image gradient information, and a rectangular building extraction result for the building is obtained. The method further comprises a step of pre-processing and a step of post-processing. Problems of a bad straight line segment extraction result, insufficient robustness and a poor search strategy existing in the traditional method are solved, the operation is simple, the operation speed is fast, and rectangular buildings of different colors, different materials and different sizes in the remote sensing image can be extracted accurately and high efficiently.
Description
Technical field
The invention belongs to Remote Sensing Target recognition technology field, be specifically related to a kind of Rectangle building extracting method based on high-resolution remote sensing image and system.
Background technology
High-resolution remote sensing image is the important information source of basic geographic database, and buildings, as the important man-made features target of a wherein class, realizes it and automatically extracts significant.Because building roof material is varied, the spectral signature shown in remote sensing images varies, and adopts the remote sensing image classification extracting method based on spectrum to be difficult to obtain desirable effect.And most buildings has regular shape, edge is straight line, there is the distinctive geometric properties such as right angle flex point, therefore, geometric properties is the Main Basis that buildings extracts.At present, Chinese scholars has proposed the method that multiple applicating geometric feature carries out buildings extraction:
Inscription on pottery soldiers etc. (2003) adopt the method for geometry meta analysis, after extracting the straight-line segment in remote sensing images, search can form the basic structural element of rectangle, and carry out merging to structural element and rebuild, automatically from the image of aviation city, extract Rectangle building [inscription on pottery soldier, Tian Yan, Zhang Jun, Deng. aerial image Rectangle building extraction method research [J]. aerospace journal, 2003,24 (4): 341-347.].Cui etc. (2012) propose the algorithm that a kind of geometric properties of complex and gray feature carry out buildings extraction.The method is according to the characteristic of spatial distribution of buildings and Hough transform characteristic, at screening edge, Hough transform territory line segment, extract potential object edge line segment, again according to the gray feature on roof, finally extract buildings [Cui S exactly, Yan Q, Reinartz P.Complex building description and extraction based on Hough transformation and cycle detection [J] .Remote Sensing Letters, 2012,3 (2): 151-159.].M.Izadi and P.Saeedi(2012) by Burns algorithm detected image cathetus feature, according to geometry feature, polygon marshalling is carried out to straight-line segment and connect extraction building roof, and utilize shade assisted reconstruction three-dimensional building model [M.Izadi and P.Saeedi, Three-Dimensional Polygonal Building Model Estimation From Single Satellite Images [J], Geoscience and Remote Sensing, IEEE Transactions on, 2012,50:2254-2272.].
The analysis found that, existing algorithm is all on the basis of line segment extraction, and the strategy combination of being organized into groups by consciousness candidate region, recycling priori differentiate whether extract region be buildings.But these methods all have some limitations, and are mainly manifested in: 1. line segment extraction less effective, not only computation complexity is high for traditional line segment extraction algorithm, and under complicated earth surface environment, effect is poor; 2. Algorithm robustness is not enough, and classic method is directly organized into groups straight-line segment, is easily subject to noise, and in the remote sensing images of scene complexity, marshalling result is undesirable; 3. rectangular search strategy imperfection, classic method often directly abandons L-type, the U-shaped structure finally failing to form complete rectangular, and this will cause holiday building target.
Summary of the invention
The invention provides a kind of Rectangle building extracting method based on high-resolution remote sensing image and system, object is the deficiency overcoming the existence of existing method, realizes extracting Rectangle building efficiently and accurately.
In order to solve the problems of the technologies described above, the technical scheme that the present invention proposes is as follows:
Based on a Rectangle building extracting method for high-resolution remote sensing image, its step comprises:
1) extract the straight-line segment feature in high-resolution remote sensing image, then according to the spatial relationship between straight-line segment, the straight-line segment adjacent in space, angle is similar to is fitted to new straight-line segment;
2) utilize line segment extraction result to extract right-angle structure, then organize right-angle structure to form rectangular configuration element;
3) being combined into the isolated right-angle structure of complete rectangular structure and U-shaped structure for failing, utilizing image gradient information to supplement undrawn border, the rectangular configuration obtaining buildings extracts result.
Further, in said method, step 1), before the described straight-line segment feature of extraction, carries out pre-service to remote sensing images, with smoothed image noise, and Retain edge information; Also comprise post-processing step after step 3): reject the false Rectangle building target of part, merge multiple rectangles overlapping on same roof.
Illustrate each step of the inventive method below:
(1) Image semantic classification step;
Adopt two-sided filter to carry out pre-service to remote sensing images in this step, two-sided filter has while carrying out filtering noise reduction, keeps the feature of image edge information, contributes to extracting buildings edge line section better.
Two-sided filter, except the parameter c determined by space length, also add the spectrum parameter s determined by pixel grey scale difference, and make pixel value differ less pixel and have larger filtering weighting, function is as follows:
X in formula, ε is volume coordinate, function h (x) represents x place pixel value after filtering, f (x) represents original image x place pixel value, and c (ε, x) represents the weight relevant to differences in spatial location, s (f (ε), f (x)) represent the weight coefficient with pixel grey scale difference correlation, k (x) is normalized weight coefficient, and computing method are as follows:
(2) line segment detection step;
Straight-line segment is the basis generating rectangular configuration, and the extraction effect of straight-line segment will determine the extraction accuracy of rectangular configuration.
In this step, first the straight-line segment feature in LSD line segment extraction algorithm extraction image is utilized, LSD(Line Segment Detector) algorithm be by the people such as Rafael G in 2010 propose, in linear session, the rectilinearity feature in image can be extracted rapidly and accurately.Then according to the spatial relationship between straight-line segment, the straight-line segment that adjacent in space, angle are similar to is fitted to new straight-line segment.As shown in Figure 2, according to the space length BC between line segment AB, CD and angle theta 3, by threshold determination AB and CD whether each other extensible line segment.If satisfy condition, utilize summit A, B, C, D of least square fitting line segment, obtain line segment A
1b
1as the result that short-term merges.
(3) rectangular configuration generation step;
First according to spatial relationship, straight-line segment is formed right-angle structure, then combine further based on right-angle structure, two right-angle structures are formed U-shaped structure, then by two U-shaped structures or a U-shaped structure and a right-angle structure composition rectangular configuration element.Comprise following sub-step:
1) right-angle structure extracts
According to the spatial relationship between line segment, Distance geometry angle threshold search is utilized to go out likely to be formed the line segment pair at right angle; Then interrupt the line segment of T font and cross combination in point of intersection, then suppress according to interrupting rear line segment length, ignore comparatively short line segment, generate corresponding right-angle structure element.
Because the shade of buildings also may comprise right angle feature, for reducing the interference that shade extracts building with rectangular, the false right-angle structure formed by shade is needed to delete.For detecting the profile whether right-angle structure is shade, for every bar right-angle side sets up the buffer zone (as shown in Figure 3) of a fixed width inside right-angle structure, differentiate whether this right angle element is shadow outline by the grey scale pixel value comprised in buffer zone, shadow outline, by deleted, is not used in the extraction of rectangle.
2) U-shaped structure extraction
U-shaped structure is made up of two right-angle structures, is the intermediate structure element of composition rectangle.Travel through each right-angle structure element, according to the most suitable right-angle structure element of following criterion Stepwise Screening in its neighborhood, merge with it and form U-shaped structure:
A. two structural element spacings are less than threshold value, and have a common edge (common edge, refer to each limit in adjacent two structural elements, the angle between them is less than threshold value and distance is less than threshold value);
B. always open direction is not identical;
C. in the quadrant that current right-angle structure L two limits surround;
D. above condition is met and two right-angle structure right-angled apices are nearest.
Merge two optimum matching right-angle structures and generate U-shaped structure: retain not common limit, as U-shaped two sides; Two right-angled apices are connected as U-shaped base.
3) rectangular configuration is extracted
After obtaining U-shaped structure, U-shaped structure or right-angle structure element merge needs with other further, composition rectangular configuration.
The method merging two U-shaped structures is each U-shaped structure of traversal, searches for the U-shaped structural element of the best that can merge with it in its proximity space according to the following rules:
A. two structural element spacings are less than threshold value, and two sides are all common edge;
B. two U-shaped opening directions are contrary;
C. another U-shaped base is in the side of the opening direction on current U-shaped base;
D. above condition is met, and nearest with current U-shaped structure base.
Merge the U-shaped structure of two optimum matching: using four summits of two U-shaped four right-angled apices as rectangle.To the U-shaped structure of failing to find coupling U-shaped, in its proximity space, search for the best right-angle structure that can merge with it according to the following rules:
A. two structural element spacings are less than threshold value, and have a common edge;
B. conllinear limit is not U-shaped structure base;
C. right-angle structure and U-shaped structure other summits except common edge summit are in the same side of common edge;
D. the common edge bearing of trend of right-angle structure is contrary with the common edge bearing of trend of U-shaped structure;
E. meet above condition, and right-angled apices and U-shaped base nearest.
Merge the right-angle structure of current U-shaped structure and optimum matching, U-shaped two base angles and right-angle structure right-angled apices form three summits of rectangle, and the 4th summit is the intersection point not belonging to the right-angle side of common edge in the U-shaped side not belonging to common edge and right angle.
(4) Elastic forming board search step
Being combined into the isolated right-angle structure of complete rectangular structure and U-shaped structure for failing, setting up Rectangular Elastic skeleton pattern, utilize image gradient information to supplement undrawn border, complete the extraction of rectangular configuration.So-called Rectangular Elastic profile refers to a rectangular profile having segment boundary position undetermined, unknown boundary can change in certain scope, as shown in Figure 4, solid line represents the border detected, dotted line represents border to be determined, unknown boundary changes within the specific limits, and its optimum position will be determined according to image gradient information further, forms final rectangular profile.
If the image gradient of unknown boundary is f (x),
x is the unknown boundary line segment in image, and p is the point on unknown boundary, and g (p) is a p place pixel gradient value, and n is the number of pixels comprised in x, uses Sobel operator (L.S.Davis, 1975) as image gradient operator in research.
Unknown boundary actual position x
0a maximum value of corresponding f (x), and the image gradient of actual position is significantly higher than the image gradient of other positions before herein, i.e. f (x
0) >>f (x), x<x
0.
In formula, λ is scale-up factor, and λ larger expression conspicuousness requires higher.Similarly, two unknown boundary defining method for right-angle structure L are as follows,
for bivector (a, b),
for bivector (a ', b '),
represent a ' <a and b ' <b:
(5) post-processing step.
Consider that road or other non-fabric structures may extract the excessive rectangle of length breadth ratio, first according to the length breadth ratio generating rectangle, reject the false Rectangle building target of part.
In addition, the roof structure of part buildings is comparatively complicated, will extract multiple rectangular configuration by rectangular extraction algorithm from image.Multiple rectangular configuration of a building roof finally will merge by this technology, and concrete scheme is as follows:
1) set up a non-directed graph, using each rectangular configuration as a node, the overlapping relation between rectangle is as the limit between node;
2) connected components all in figure is asked;
3) the minimum outsourcing rectangle that the polygon in each connected component combines is tried to achieve, as the rectangular target after merging.
Based on a Rectangle building extraction system for high-resolution remote sensing image, it comprises:
Straight-line detection module, for extracting the straight-line segment feature in high-resolution remote sensing image, and according to the spatial relationship between straight-line segment, fits to new straight-line segment by the straight-line segment adjacent in image, angle is similar to;
Right-angle structure marshalling module, connects described straight-line detection module, for utilizing line segment extraction result to extract right-angle structure, by close for position in image, orthogonal sets of line segments synthesis right-angle structure;
Rectangle building generation module, connect described right-angle structure marshalling module, for by orthogonal for right-angle structure element combinations structure, the isolated right-angle structure of complete rectangular structure and U-shaped structure is combined into failing, utilize image gradient information to supplement undrawn border, the rectangular configuration obtaining buildings extracts result.
Further, said system also comprises: pretreatment module, for before the described straight-line segment feature of extraction, carries out pre-service, and pretreated image is inputted described line segment detection module to remote sensing images; Post-processing module, for rejecting the false Rectangle building target of part, and merges multiple rectangles overlapping in same roof.
The present invention is towards complicated earth surface background, based on high-resolution remote sensing image, by line segment detection algorithm, more outstanding straight-line segment organize into groups thought and more intelligent elasticity search strategy more accurately, propose a kind of Rectangle building extracting method, solve classic method cathetus section and extract the problem that result is bad, robustness is not enough and search strategy is not good enough, effectively can extract the Rectangle building of different colours in remote sensing images, unlike material, different size, simple to operate, fast operation, has higher practical value.
Accompanying drawing explanation
Fig. 1 is steps flow chart and the module structure drafting of buildings extracting method of the present invention.
Fig. 2 is spatial relationship between two straight-line segments and fitting result schematic diagram.
Fig. 3 is right-angle structure buffer zone schematic diagram.
Fig. 4 is the Rectangular Elastic skeleton pattern schematic diagram generated by U-shaped structure and right-angle structure.
Fig. 5 is the high-resolution remote sensing image in somewhere, Shenyang, Liaoning.
Fig. 6 is the result that application LSD algorithm extracts straight-line segment.
Fig. 7 is the connection result of adjacent segments.
Fig. 8 is the extraction result of right-angle structure.
Fig. 9 is the extraction result of rectangular configuration.
Figure 10 is the net result that application the inventive method extracts building with rectangular.
Figure 11 is the module composition diagram of the Rectangle building extraction system of application the inventive method.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further detail.
Rectangle building extracting method based on high-resolution remote sensing image of the present invention, its flow process as shown in Figure 1.Buildings extracting method of the present invention and system comprise data and read in module, read in remote sensing images to be extracted; Image pre-processing module, carries out the operation such as filtering, denoising to remote sensing images, improves the contrast of target and other atural objects; Straight-line detection module, uses line detection algorithm to extract linear feature rapidly and accurately; Right-angle structure marshalling module and Rectangle building generation module, adopt the method based on right-angle structure Perception Agent to carry out building target extraction.
Example: for a panel height resolution remote sensing images in somewhere, Shenyang City, Liaoning Province, application the inventive method carries out Rectangle building extraction to it.This image is Panchromatic image, and spatial resolution is 2m.
1. Image semantic classification
The remote sensing images of input as shown in Figure 5.Carry out Image semantic classification to remote sensing images, utilize bilateral filtering algorithm, smoothed image noise, weakens texture variations, Retain edge information.
2. line segment detection
First apply the straight-line segment in LSD algorithm extraction image, result as shown in Figure 6.
Next the preliminary straight-line segment extracted is connected, the adjacent and sets of line segments that direction is close of matching.Setting space distance threshold is 3 pixels, and angle similar threshold value is π/16, searches for all sets of line segments that can be connected, then each group line segment is carried out matching, replaces initial segment, obtain final line segment extraction result as shown in Figure 7 by the result of matching.Wherein, word space threshold and angle similar threshold value be desirable different value in other embodiments, and preferable range is: word space threshold 1 ~ 5 pixel, angle similar threshold value π/32 ~ π/12.
3. rectangular configuration generates
First, line segment extraction result is utilized to extract right-angle structure.Between setting line segment, word space threshold is 3 pixels, and between line segment, angle threshold value is (7 π/16,9 π/16), and line segment length threshold value is 5 pixels, and the extraction result of right-angle structure as shown in Figure 8.Then right-angle structure is organized to form rectangular configuration element.Above-mentioned each threshold value is desirable different value in other embodiments, and preferable range is: word space threshold 1 ~ 5 pixel between line segment, π/32, angle threshold value 15 π/32 ~ 19 between line segment, line segment length threshold value 3 ~ 10 pixels.
4. Elastic forming board search
Rectangular Elastic skeleton pattern is set up to the right-angle structure and U-shaped structure of failing to form complete rectangular structure: in the Rectangular Elastic skeleton pattern of U-shaped structural generation, the variation range of unknown boundary is the half from shorter lateral sides length, to compared with 1.5 times of long side lengths; The unknown boundary variation range of Rectangular Elastic skeleton pattern that right-angle structure generates is that it adjacently separately determines 0.5 times of boundary length to 1.5 times; The value setting edge undetermined conspicuousness coefficient lambda is 1.25.The rectangle of application elasticity profile model generation and the rectangle result of application right-angle structure composition are as shown in Figure 9.Wherein edge undetermined conspicuousness coefficient lambda desirable different value in other embodiments, preferred span is 1 ~ 1.5.
5. rectangular extraction aftertreatment
Setting rectangular aspect ratio threshold value is 20, rejects the interference rectangle that length breadth ratio exceedes threshold value; Then merged by the rectangle intersected, final Rectangle building extracts result as shown in Figure 10.
Figure 11 is the module composition diagram of the Rectangle building extraction system based on high-resolution remote sensing image of the employing said method of the present embodiment, specifically comprises:
Pretreatment module, for before the described straight-line segment feature of extraction, carries out pre-service to remote sensing images, with smoothed image noise, and Retain edge information;
Straight-line detection module, connects described pretreatment module, for extracting the straight-line segment feature in high-resolution remote sensing image, and according to the spatial relationship between straight-line segment, the straight-line segment adjacent in image, angle is similar to is fitted to new straight-line segment;
Right-angle structure marshalling module, connects described straight-line detection module, for utilizing line segment extraction result to extract right-angle structure, by close for position in image, orthogonal sets of line segments synthesis right-angle structure;
Rectangle building extraction module, connect described right-angle structure marshalling module, rectangular configuration is formed for organizing right-angle structure, the isolated right-angle structure of complete rectangular structure and U-shaped structure is combined into failing, utilize image gradient information to supplement undrawn border, the rectangular configuration obtaining buildings extracts result;
Post-processing module, connects described Rectangle building generation module, for rejecting the false Rectangle building target of part, and merges multiple rectangles overlapping in same roof.
Above embodiment is only in order to illustrate technical scheme of the present invention but not to be limited; those of ordinary skill in the art can modify to technical scheme of the present invention or equivalent replacement; and not departing from the spirit and scope of the present invention, protection scope of the present invention should be as the criterion with described in claim.
Claims (10)
1., based on a Rectangle building extracting method for high-resolution remote sensing image, its step comprises:
1) extract the straight-line segment feature in high-resolution remote sensing image, then according to the spatial relationship between straight-line segment, the straight-line segment adjacent in image, angle is similar to is fitted to new straight-line segment;
2) utilize line segment extraction result to extract right-angle structure, then organize right-angle structure to form rectangular configuration element;
3) being combined into the isolated right-angle structure of complete rectangular structure and U-shaped structure for failing, utilizing image gradient information to supplement undrawn border, the rectangular configuration obtaining buildings extracts result.
2. the method for claim 1, is characterized in that: step 1), before the described straight-line segment feature of extraction, carries out pre-service to remote sensing images, with smoothed image noise, and Retain edge information; Also comprise post-processing step after step 3): reject the false Rectangle building target of part, merge multiple rectangles overlapping on same roof.
3. method as claimed in claim 2, is characterized in that, adopts two-sided filter to carry out described pre-service.
4. the method for claim 1, is characterized in that: step 1) utilizes the straight-line segment feature in LSD line segment extraction algorithm extraction image.
5. the method for claim 1, it is characterized in that: step 2) first according to spatial relationship, straight-line segment is formed right-angle structure, then combine further based on right-angle structure, two right-angle structures are formed U-shaped structure, again by two U-shaped structures or a U-shaped structure and a right-angle structure composition rectangular configuration element, specifically comprise following sub-step:
2-1) right-angle structure extracts: according to the spatial relationship between line segment, utilize Distance geometry angle threshold search to go out likely to be formed the line segment pair at right angle; Then interrupt the line segment of T font and cross combination in point of intersection, then suppress according to interrupting rear line segment length, ignore comparatively short line segment, generate corresponding right-angle structure element;
2-2) U-shaped structure extraction: U-shaped structure is made up of two right-angle structures, it is the intermediate structure element of composition rectangle, travel through each right-angle structure element, according to the most suitable right-angle structure element of following criterion Stepwise Screening in its neighborhood, merge with it and form U-shaped structure:
A. two structural element spacings are less than threshold value, and have a common edge, and described common edge refers to each limit in adjacent two structural elements, and the angle between them is less than threshold value and distance is less than threshold value,
B. always open direction is not identical,
C. in the quadrant that current right-angle structure L two limits surround,
D. above condition is met and two right-angle structure right-angled apices are nearest;
Merge two optimum matching right-angle structures and generate U-shaped structure: retain not common limit, as U-shaped two sides; Two right-angled apices are connected as U-shaped base;
2-3) rectangular configuration is extracted: after obtaining U-shaped structure, U-shaped structure or right-angle structure element merge with other further, composition rectangular configuration, the method merging two U-shaped structures is each U-shaped structure of traversal, searches for the U-shaped structural element of the best that can merge with it in its proximity space according to the following rules:
A. two structural element spacings are less than threshold value, and two sides are all common edge,
B. two U-shaped opening directions are contrary,
C. another U-shaped base is in the side of the opening direction on current U-shaped base,
D. above condition is met, and nearest with current U-shaped structure base;
Merge the U-shaped structure of two optimum matching: using two U-shaped four right-angled apices as four summits of rectangle, to the U-shaped structure of failing to find coupling U-shaped, in its proximity space, search for the best right-angle structure that can merge with it according to the following rules:
A. two structural element spacings are less than threshold value, and have a common edge,
B. conllinear limit is not U-shaped structure base,
C. right-angle structure and U-shaped structure other summits except common edge summit are in the same side of common edge,
D. the common edge bearing of trend of right-angle structure is contrary with the common edge bearing of trend of U-shaped structure,
E. meet above condition, and right-angled apices and U-shaped base nearest;
Merge the right-angle structure of current U-shaped structure and optimum matching, U-shaped two base angles and right-angle structure right-angled apices form three summits of rectangle, and the 4th summit is the intersection point not belonging to the right-angle side of common edge in the U-shaped side not belonging to common edge and right angle.
6. method as claimed in claim 5, is characterized in that: step 2-1) when extracting right-angle structure, for the building effects comprising right angle feature, the false right-angle structure formed by shade is deleted, to reduce the interference that shade extracts building with rectangular; For detecting the profile whether right-angle structure is shade, for every bar right-angle side sets up the buffer zone of a fixed width inside right-angle structure, differentiate whether this right angle element is shadow outline by the grey scale pixel value comprised in buffer zone.
7. the method for claim 1, it is characterized in that: step 3) adopts Rectangular Elastic profile to realize the described image gradient information that utilizes and supplements undrawn border, so-called Rectangular Elastic profile refers to a rectangular profile having segment boundary position undetermined, unknown boundary can change in certain scope, its optimum position is determined according to image gradient information, forms final rectangular profile.
8. the method for claim 1, is characterized in that, multiple rectangular configuration of extract building roof is merged, and obtains final Rectangle building and extracts result; The method of described merging is as follows:
A) set up a non-directed graph, using each rectangular configuration as a node, the overlapping relation between rectangle is as the limit between node;
B) connected components all in figure is asked;
C) the minimum outsourcing rectangle that the polygon in each connected component combines is tried to achieve, as the rectangular target after merging.
9. adopt the Rectangle building extraction system based on high-resolution remote sensing image of method described in claim 1, it is characterized in that, comprising:
Straight-line detection module, for extracting the straight-line segment feature in high-resolution remote sensing image, and according to the spatial relationship between straight-line segment, fits to new straight-line segment by the straight-line segment adjacent in image, angle is similar to;
Right-angle structure marshalling module, connects described straight-line detection module, for utilizing line segment extraction result to extract right-angle structure, by close for position in image, orthogonal sets of line segments synthesis right-angle structure;
Rectangle building generation module, connect described right-angle structure marshalling module, for by orthogonal for right-angle structure element combinations structure, the isolated right-angle structure of complete rectangular structure and U-shaped structure is combined into failing, utilize image gradient information to supplement undrawn border, the rectangular configuration obtaining buildings extracts result.
10. system as claimed in claim 9, is characterized in that, also comprise: pretreatment module, for before the described straight-line segment feature of extraction, carries out pre-service, and pretreated image is inputted described line segment detection module to remote sensing images; Post-processing module, for rejecting the false Rectangle building target of part, and merges multiple rectangles overlapping in same roof.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410093123.4A CN104915672B (en) | 2014-03-13 | 2014-03-13 | A kind of Rectangle building extracting method and system based on high-resolution remote sensing image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410093123.4A CN104915672B (en) | 2014-03-13 | 2014-03-13 | A kind of Rectangle building extracting method and system based on high-resolution remote sensing image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104915672A true CN104915672A (en) | 2015-09-16 |
CN104915672B CN104915672B (en) | 2018-08-03 |
Family
ID=54084722
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410093123.4A Expired - Fee Related CN104915672B (en) | 2014-03-13 | 2014-03-13 | A kind of Rectangle building extracting method and system based on high-resolution remote sensing image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104915672B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105574521A (en) * | 2016-02-25 | 2016-05-11 | 民政部国家减灾中心 | House contour extraction method and apparatus thereof |
CN105719306A (en) * | 2016-01-26 | 2016-06-29 | 郑州恒正电子科技有限公司 | Rapid building extraction method from high-resolution remote sensing image |
CN106650749A (en) * | 2016-11-16 | 2017-05-10 | 桂林理工大学 | Method for plotting right-angled building in high-resolution optical image |
CN107092877A (en) * | 2017-04-12 | 2017-08-25 | 武汉大学 | Remote sensing image roof contour extracting method based on basement bottom of the building vector |
CN107369150A (en) * | 2016-05-12 | 2017-11-21 | 株式会社理光 | A kind of method and rectangular target detection means for detecting rectangular target |
WO2019148902A1 (en) * | 2018-02-02 | 2019-08-08 | 北京三快在线科技有限公司 | Polygonal region detection |
CN110390267A (en) * | 2019-06-25 | 2019-10-29 | 东南大学 | A kind of mountain landscape Building extraction method and apparatus based on high score remote sensing image |
CN111325684A (en) * | 2020-02-01 | 2020-06-23 | 武汉大学 | Semi-automatic high-spatial-resolution remote sensing image extraction method for buildings with different shapes |
CN112348836A (en) * | 2020-11-06 | 2021-02-09 | 二十一世纪空间技术应用股份有限公司 | Method and device for automatically extracting building outline |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113792728A (en) * | 2021-08-06 | 2021-12-14 | 南宁学院 | High-precision visual positioning method |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0626654A2 (en) * | 1993-05-25 | 1994-11-30 | Matsushita Electric Industrial Co., Ltd. | Apparatus for measuring intervehicle distance by stereo vision |
CN101976349A (en) * | 2010-10-21 | 2011-02-16 | 西北工业大学 | Hough-transform-based method for identifying transportation junction targets in remote-sensing image |
CN102122344A (en) * | 2011-01-07 | 2011-07-13 | 南京理工大学 | Road border detection method based on infrared image |
CN102592454A (en) * | 2012-02-29 | 2012-07-18 | 北京航空航天大学 | Intersection vehicle movement parameter measuring method based on detection of vehicle side face and road intersection line |
CN102938066A (en) * | 2012-12-07 | 2013-02-20 | 南京大学 | Method for reconstructing outer outline polygon of building based on multivariate data |
-
2014
- 2014-03-13 CN CN201410093123.4A patent/CN104915672B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0626654A2 (en) * | 1993-05-25 | 1994-11-30 | Matsushita Electric Industrial Co., Ltd. | Apparatus for measuring intervehicle distance by stereo vision |
CN101976349A (en) * | 2010-10-21 | 2011-02-16 | 西北工业大学 | Hough-transform-based method for identifying transportation junction targets in remote-sensing image |
CN102122344A (en) * | 2011-01-07 | 2011-07-13 | 南京理工大学 | Road border detection method based on infrared image |
CN102592454A (en) * | 2012-02-29 | 2012-07-18 | 北京航空航天大学 | Intersection vehicle movement parameter measuring method based on detection of vehicle side face and road intersection line |
CN102938066A (en) * | 2012-12-07 | 2013-02-20 | 南京大学 | Method for reconstructing outer outline polygon of building based on multivariate data |
Non-Patent Citations (3)
Title |
---|
王波: "一种改进的Hough变换在建筑物提取中的算法研究", 《城市勘测》 * |
韩丽荣: "复杂场景图像中建筑物提取技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
魏志强: "数字城市中矩形建筑物区域的自动获取", 《计算机科学》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105719306B (en) * | 2016-01-26 | 2018-09-11 | 郑州恒正电子科技有限公司 | A kind of building rapid extracting method in high-resolution remote sensing image |
CN105719306A (en) * | 2016-01-26 | 2016-06-29 | 郑州恒正电子科技有限公司 | Rapid building extraction method from high-resolution remote sensing image |
CN105574521A (en) * | 2016-02-25 | 2016-05-11 | 民政部国家减灾中心 | House contour extraction method and apparatus thereof |
CN105574521B (en) * | 2016-02-25 | 2019-01-29 | 民政部国家减灾中心 | Outline of house extracting method and device |
CN107369150B (en) * | 2016-05-12 | 2020-10-27 | 株式会社理光 | Method for detecting rectangular target and rectangular target detection device |
CN107369150A (en) * | 2016-05-12 | 2017-11-21 | 株式会社理光 | A kind of method and rectangular target detection means for detecting rectangular target |
CN106650749B (en) * | 2016-11-16 | 2019-10-11 | 桂林理工大学 | The drawing method of right angle building in a kind of high-resolution optical image |
CN106650749A (en) * | 2016-11-16 | 2017-05-10 | 桂林理工大学 | Method for plotting right-angled building in high-resolution optical image |
CN107092877A (en) * | 2017-04-12 | 2017-08-25 | 武汉大学 | Remote sensing image roof contour extracting method based on basement bottom of the building vector |
CN107092877B (en) * | 2017-04-12 | 2020-01-24 | 武汉大学 | Remote sensing image roof contour extraction method based on building base vector |
WO2019148902A1 (en) * | 2018-02-02 | 2019-08-08 | 北京三快在线科技有限公司 | Polygonal region detection |
US11308710B2 (en) | 2018-02-02 | 2022-04-19 | Beijing Sankuai Online Technology Co., Ltd | Polygonal region detection |
CN110390267A (en) * | 2019-06-25 | 2019-10-29 | 东南大学 | A kind of mountain landscape Building extraction method and apparatus based on high score remote sensing image |
CN110390267B (en) * | 2019-06-25 | 2021-06-01 | 东南大学 | Mountain landscape building extraction method and device based on high-resolution remote sensing image |
CN111325684A (en) * | 2020-02-01 | 2020-06-23 | 武汉大学 | Semi-automatic high-spatial-resolution remote sensing image extraction method for buildings with different shapes |
CN112348836A (en) * | 2020-11-06 | 2021-02-09 | 二十一世纪空间技术应用股份有限公司 | Method and device for automatically extracting building outline |
CN112348836B (en) * | 2020-11-06 | 2024-03-12 | 二十一世纪空间技术应用股份有限公司 | Method and device for automatically extracting building outline |
Also Published As
Publication number | Publication date |
---|---|
CN104915672B (en) | 2018-08-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104915672A (en) | Rectangular building extraction method and system based on high-resolution remote sensing image | |
CN102938066B (en) | A kind of based on the polygonal method of multivariate data reconstruction buildings outline | |
CN105931295B (en) | A kind of geologic map Extracting Thematic Information method | |
Chen et al. | An improved license plate location method based on edge detection | |
CN106780582B (en) | The image significance detection method merged based on textural characteristics and color characteristic | |
Huang et al. | A top-down strategy for buildings extraction from complex urban scenes using airborne LiDAR point clouds | |
CN103413347A (en) | Extraction method of monocular image depth map based on foreground and background fusion | |
CN103489193B (en) | OO high-resolution remote sensing image change detecting method based on convergence strategy | |
CN101840582B (en) | Boundary digitizing method of cadastral plot | |
CN103035006A (en) | High-resolution aerial image partition method based on LEGION and under assisting of LiDAR | |
CN103095996A (en) | Multi-sensor video fusion method based on space-time conspicuousness detection | |
CN104318563A (en) | Organ skeleton extraction method based on medical images | |
Li et al. | Urban building damage detection from very high resolution imagery by One-Class SVM and shadow information | |
Zhang et al. | Road extraction and intersection detection based on tensor voting | |
Hongjian et al. | 3D building reconstruction from aerial CCD image and sparse laser sample data | |
CN106023096B (en) | Eliminate the image matching method that edge sawtoothization influences | |
Bhadauria et al. | Building extraction from satellite images | |
CN112232249A (en) | Remote sensing image change detection method and device based on depth features | |
Li et al. | Detecting building changes using multi-modal Siamese multi-task networks from very high resolution satellite images | |
CN104268836A (en) | Watershed segmentation mark point extraction method based on local area homogeneity indexes | |
CN103955935A (en) | Salient region detection method based on local self-adaptation regression | |
Kang et al. | Line feature extraction from RGB laser point cloud | |
Lei et al. | Fast Multi-Object Image Segmentation Algorithm Based on CV Model. | |
Zhang et al. | Multi-scale feature fusion for hyperspectral and LiDAR data joint classification | |
Gou et al. | 3D point cloud global denoising algorithm based on different noise characteristics |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180803 |